Plug-in micro set-top box multi-screen interactive content data adaptive processing method and system

Abstract

The application discloses a plug-in micro set-top box multi-screen interactive content data adaptive processing method and system, relates to the technical field of broadcast television; the application is specially aimed at the scene of plug-in micro set-top box multi-screen interaction, and aims to solve the content processing bottleneck caused by local resource limitation and the problem of being unable to realize multi-screen interaction; based on a star structure formed by a central computing platform and a plug-in micro set-top box, the central computing platform is centrally managed, the plug-in micro set-top box only serves as light-weight data collection, streaming media receiving and processing, the plug-in micro set-top box uploads the collected device parameters to the platform, the platform adaptively processes content to generate streaming media, realizes multi-screen interaction through real-time processing, and considering predictive processing, distribution and resource scheduling, high adaptive efficiency, seamless switching, excellent picture quality and high scalability are realized.

Description

Technical Field

[0001] This application relates to the technical field of broadcasting and television, and in particular to a method and system for adaptive processing of multi-screen interactive content data in plug-in mini set-top boxes. Background Technology

[0002] With the increasing richness of digital media content and the continuous growth in user demand for immersive and seamless experiences, multi-screen interaction technology has become an indispensable core component of modern smart living and office environments. Whether it's sharing entertainment content in smart home scenarios, personalized viewing experiences in home theater systems, or efficient information flow, all place unprecedented demands on real-time content synchronization and optimization across devices. Against this backdrop, existing technologies have failed to provide an effective strategy for managing and optimizing these cross-device pre-loaded resources, becoming a key obstacle restricting further improvements in multi-screen interaction experiences.

[0003] Therefore, how to overcome the limitations of computing and storage resources in traditional plug-in micro set-top boxes, since a plug-in micro set-top box can only control one device, and how to build a data processing method that can efficiently, in real time and seamlessly support multi-screen interaction, thereby completely eliminating the user experience damage caused by content adaptive processing, has become a key challenge and an urgent technical problem for those skilled in the art. Summary of the Invention

[0004] In order to at least overcome the above-mentioned shortcomings in the prior art, the purpose of this application is to provide a method and system for adaptive processing of multi-screen interactive content data in plug-in micro set-top boxes to solve the above problems.

[0005] In a first aspect, this application provides a method for processing multi-screen interactive content data in a plug-in mini set-top box, including:

[0006] Based on the target user's historical playback behavior, a predictive model for the target plug-in mini set-top box is obtained; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in mini set-top box is configured such that if the user streams the same playback content from one plug-in mini set-top box to another plug-in mini set-top box, then the other plug-in mini set-top box is selected as the target plug-in mini set-top box.

[0007] In response to the detection of the first user action of the target user, the current playback behavior is obtained.

[0008] Based on the current playback behavior and the prediction model, the target plug-in micro set-top box is obtained.

[0009] Based on the current playback behavior, the target streaming media data is obtained and sent to the target plug-in micro set-top box;

[0010] The target plug-in mini set-top box is also configured to: cache target streaming media data and, in response to detecting a second user action, retrieve the target streaming media data for the target user to choose to play.

[0011] In one possible implementation, a predictive model of the target plug-in set-top box is obtained based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in set-top box is configured to use the other plug-in set-top box as the target plug-in set-top box if the user transfers the same playback content from one playback plug-in set-top box to another, including:

[0012] Based on the target user's historical playback behavior, playback device data and playback content data are time-series processed to obtain user usage logs;

[0013] Based on user logs, the probability of a user switching the currently playing content to another target plug-in mini set-top box is obtained; where the switching probability represents the likelihood that a user will switch the currently playing content from the current plug-in mini set-top box to the target plug-in mini set-top box;

[0014] Construct a training dataset based on the switching probabilities;

[0015] Based on user logs and the Transformer model, a time series prediction model is obtained.

[0016] Use user usage logs as input to output the switching probability of each potential target plug-in micro set-top box;

[0017] The time series prediction model is trained based on historical playback behavior and training dataset to minimize the cross-entropy loss between the prediction probability of the time series prediction model based on historical playback behavior and the real training dataset.

[0018] The time series prediction model that minimizes cross-entropy loss is used as the prediction model for the target plug-in micro set-top box.

[0019] In one possible implementation, a predictive model for the target plug-in mini set-top box is obtained based on the target user's historical playback behavior, including:

[0020] Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the identifier of the currently playing plug-in mini set-top box; the user interaction event includes a third user action, the third user action including at least a user initiating a content playback request on any device; in response to detecting a third user action, receive the third user action and the identifier of the currently playing plug-in mini set-top box sent by the currently playing plug-in mini set-top box.

[0021] Based on the content playback request, obtain the original streaming media from the content source;

[0022] The central computing platform sends the original streaming media, based on the identifier of the currently playing plug-in mini set-top box, to the currently playing plug-in mini set-top box; where:

[0023] The currently playing plug-in mini set-top box performs adaptive content data processing based on the acquired current playback capability parameters to obtain a currently playing streaming media adapted to the currently playing plug-in mini set-top box; wherein, the adaptive content data processing includes:

[0024] When the resolution of the original streaming media is higher than the current playback capability parameters, based on the original streaming media and an image scaling algorithm, a new streaming media with a video resolution adapted to the current playback resolution is obtained.

[0025] When the encoding format of the original streaming media is not directly supported by the current playback device, it is converted according to the encoding format of the original streaming media to obtain the current playback streaming media with the encoding format supported by the current playback device;

[0026] Based on the network bandwidth of the current playback plug-in mini set-top box, and using rate-distortion theory, the highest bitrate of the currently playing streaming media is obtained.

[0027] Based on the network bandwidth of the target plug-in micro set-top box and using rate-distortion theory, the current streaming media with the highest bitrate is obtained, including:

[0028] Maximize streaming quality and minimize distortion within network bandwidth constraints;

[0029] Based on the original streaming media, compression is performed, and the pixel difference between the original streaming media and the compressed streaming media is calculated to obtain the distortion.

[0030] Based on the distortion degree and network bandwidth, and using rate-distortion theory, we obtain the rate-distortion theory.

[0031] Based on rate-distortion theory, the minimum quantization parameter is obtained through iterative bisection to obtain the current streaming media with the highest bitrate.

[0032] In one possible implementation, in response to detecting a first user action of the target user, the current playback action is obtained, including:

[0033] Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the content identifier of the currently playing streaming media, the interruption playback progress timestamp, and the identifier of the currently playing plug-in mini set-top box; the user interaction event includes a first user action, which includes at least a power-off action of the device connected to the plug-in mini set-top box turning off.

[0034] In response to the detection of the first user action, if the streaming media content being played at the time of the device power-off action has not yet finished playing, the plug-in micro set-top box will return the status information as a trigger signal for the content interruption event to the central computing platform.

[0035] In one possible implementation, based on the state information and the currently playing streaming media, and using a prediction model of the target plug-in micro set-top box, a target streaming media adapted to the target plug-in micro set-top box is obtained, including:

[0036] In response to the first user action detected by the plug-in micro set-top box, if the device is currently playing streaming media content at the time of power-off and the content has not yet finished playing, the currently playing streaming media content will be switched from the current plug-in micro set-top box to another plug-in micro set-top box for playback.

[0037] Based on the trigger signal, obtain the content identifier and the interruption playback progress timestamp;

[0038] Based on the current status information and the content identifier of the currently playing streaming media, the predicted target plug-in micro set-top box is obtained based on the prediction model of the target plug-in micro set-top box.

[0039] Based on the content identifier and the content source, obtain the original streaming media;

[0040] Based on the predicted target plug-in micro set-top box identifier, the original streaming media is pre-distributed to the target plug-in micro set-top box's buffer.

[0041] The currently playing plug-in mini set-top box obtains the current playback capability parameters of the connected device and performs adaptive content data processing to obtain the target streaming media adapted to the target plug-in mini set-top box.

[0042] In one possible implementation, pre-distribution to the buffer of the target plug-in micro set-top box includes:

[0043] At least send to the buffer of the target plug-in mini set-top box;

[0044] And or respectively sent to the buffers of all plug-in mini set-top boxes under the home address where the target plug-in mini set-top box is located.

[0045] In one possible implementation, in response to detecting a second user action, the target streaming media is retrieved for the user to choose to play, including:

[0046] Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the content identifier of the currently playing streaming media, the interruption playback progress timestamp, and the identifier of the currently playing plug-in mini set-top box; the user interaction events include a second user action, which includes at least the power-on action of any device.

[0047] When the original media stream is obtained based on the content identifier, it is pre-distributed to the buffer of the target plug-in mini set-top box;

[0048] The target plug-in micro set-top box receives the raw streaming media distributed from the central computing platform, performs adaptive content data processing to obtain the target media stream, stores it in its local buffer, and establishes a preload queue.

[0049] The preload queue includes one or more target streaming media entries waiting in the queue. Each target streaming media entry includes the content identifier of the corresponding target streaming media, the start timestamp of preloading, and a corresponding target streaming media stored in the buffer area of ​​the target plug-in micro set-top box.

[0050] The number of target streaming media entries waiting in the preload queue does not exceed a preset maximum value. If the preset maximum value is reached, the preload queue is full.

[0051] When a new target streaming media is received and the preload queue is full, the target plug-in mini set-top box processes it according to a preset replacement strategy.

[0052] In response to the detection of a second user action, a command to retrieve the target streaming media is sent to the target plug-in mini set-top box. The target plug-in mini set-top box immediately transmits the preload queue to the target device connected to it for the user to select and play.

[0053] In one possible implementation, when a new target streaming media is received and the preload queue is full, the target plug-in micro set-top box processes the data according to a preset replacement strategy, including:

[0054] The replacement strategy is to start an independent timer for each target streaming media in the preload queue;

[0055] If a target streaming media in a preloaded queue is not accessed by a user within a preset time limit after being received, or if no deletion instruction is received from the central computing platform, the target plug-in micro set-top box will automatically delete the queue and the corresponding data stored in the buffer.

[0056] In one possible implementation, the target plug-in micro set-top box is further configured to: cache target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for optional playback by the target user, and further include:

[0057] When a user selects the target streaming media for a previously interrupted content event on the predicted target plug-in mini set-top box, the target plug-in mini set-top box reads data from the buffer and immediately resumes playback from the interrupted playback progress timestamp.

[0058] When the target plug-in micro set-top box successfully resumes playback, it returns a playback resume success signal to the central computing platform. The playback resume success signal includes at least the content identifier of the target streaming media and the identifier of the target plug-in micro set-top box.

[0059] After receiving the successful playback resume signal, the central computing platform generates a preloaded content deletion instruction and sends it to all target plug-in micro set-top boxes except those that are resuming playback. The target plug-in micro set-top boxes except those that are resuming playback are configured to have target streaming media entries in their local preload queues, which include the content identifier of the target streaming media for which playback is resuming.

[0060] Other target plug-in micro set-top boxes that receive a preload content deletion instruction immediately delete the corresponding target streaming media entry and its data in the buffer from their preload queue.

[0061] Secondly, this application provides a plug-in micro set-top box multi-screen interactive content data processing system, including an electrically connected prediction unit, a multi-screen interactive unit, and a content data processing unit;

[0062] The prediction unit is configured to: obtain a prediction model of the target plug-in mini set-top box based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in mini set-top box is configured to select the other plug-in mini set-top box as the target plug-in mini set-top box if the user streams the same playback content from one playback plug-in mini set-top box to another plug-in mini set-top box.

[0063] The multi-screen interaction unit is configured to: obtain the current playback behavior in response to the detection of the first user behavior of the target user.

[0064] The content data processing unit is configured to: obtain the target plug-in micro set-top box based on the current playback behavior and a prediction model; wherein: the target streaming media data is obtained based on the current playback behavior and sent to the target plug-in micro set-top box.

[0065] The multi-screen interaction unit is also configured to: cache target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for the target user to choose to play.

[0066] In summary, the beneficial effects that this application can achieve are:

[0067] This application proposes a method and system for adaptive processing of multi-screen interactive content data in a plug-in micro set-top box. Through a central computing platform, it realizes state information analysis, prediction and pre-distribution, and resource scheduling optimization. The set-top box performs streaming media adaptation, which solves the problems of high processing latency and inability to continue watching across home addresses in traditional solutions. It achieves the advantages of improving processing efficiency in high-concurrency scenarios with multiple devices, reducing content switching latency, optimizing resource utilization, and enhancing user experience. Attached Figure Description

[0068] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0069] Figure 1 This is a schematic diagram of the method steps in an embodiment of this application;

[0070] Figure 2 This is a schematic diagram of a star-shaped structure consisting of a central computing platform and a plug-in micro set-top box, according to an embodiment of this application.

[0071] Figure 3 This is a schematic diagram of the system structure according to an embodiment of this application. Detailed Implementation

[0072] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0073] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0074] Example 1

[0075] Please refer to the following: Figure 1 This is a schematic diagram of the steps of the adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box provided in the embodiment of the present invention. Further, the adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box may specifically include the content described in steps S1 to S32.

[0076] Step S1: Obtain a prediction model for the target plug-in mini set-top box based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in mini set-top box is configured such that if the user streams the same playback content from one plug-in mini set-top box to another plug-in mini set-top box, then the other plug-in mini set-top box is selected as the target plug-in mini set-top box.

[0077] Step S2: In response to detecting the first user action of the target user, obtain the current playback behavior.

[0078] Step S3: Based on the current playback behavior and the prediction model, obtain the target plug-in micro set-top box.

[0079] Step S31: Based on the current playback behavior, obtain the target streaming media data and send it to the target plug-in micro set-top box.

[0080] Step S32: The target plug-in micro set-top box is further configured to: cache target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for the target user to choose to play.

[0081] Because traditional plug-in mini set-top boxes cannot handle local content adaptive processing tasks in multi-screen interaction scenarios, plug-in mini set-top boxes only control one device and face problems such as insufficient computing resources. When users try to switch the playback content from the TV to the tablet, they cannot achieve multi-screen interaction due to limitations in local hardware performance.

[0082] With the increasing demand for multiple rooms and devices in real-world homes, the one-to-one model is no longer sufficient to meet real-time requirements, becoming a key bottleneck restricting the improvement of user experience. Therefore, in the implementation of this application, through analysis, it was found that transferring content processing tasks from resource-constrained plug-in micro set-top boxes to a cloud platform as a central computing platform can effectively overcome the limitations of limited local computing power. Furthermore, since user switching behavior is predictable, the original media stream is pre-distributed to potential target plug-in micro set-top boxes. The target plug-in micro set-top boxes then perform adaptive processing of the original media stream content data into streaming media adapted to the corresponding connected devices, which can shorten the response time. Based on this, a central computing platform and plug-in micro set-top box architecture is proposed. A predictive model for the target plug-in micro set-top box is established to pre-distribute streaming media and perform adaptive processing to adapt to different devices. Then, a pre-loading queue is established, and the central computing platform resources are monitored to adjust and schedule tasks.

[0083] In this embodiment, the user sets the TV to be compatible with 4K streaming media. When the user initiates a request to play the documentary "XXX" on the TV, the central computing platform receives this user interaction event. In response to detecting this user behavior, the plug-in mini set-top box connected to the TV sends its identifier and content playback request to the central computing platform. The central computing platform queries the device database to obtain the H265 encoding format and network bandwidth parameters supported by the TV. Based on the content identifiers of various streaming media in the content source, it finds the content identifier of the documentary "XXX" as Documentary A, obtains the original streaming media of Documentary A in the content source, and returns the original streaming media to the plug-in mini set-top box. The plug-in mini set-top box performs resolution conversion and transcoding on the original streaming media, and generates the highest bitrate current playback streaming media, limited by network bandwidth.

[0084] Then, when the user turns off the currently playing device, the TV, the central computing platform receives this user interaction event, takes the power-off action as the first user action, and in response to detecting this first user action, determines whether the currently playing streaming media content at the time of the power-off action has finished playing; if the content has not finished playing, the plug-in micro set-top box returns the status information as a trigger signal for the content interruption event to the central computing platform.

[0085] The central computing platform then obtains the target plug-in micro set-top box based on the prediction model of the target plug-in micro set-top box; then it distributes the original streaming media to the target plug-in micro set-top box, which adaptively processes the streaming media to match the target streaming media of the device connected to the target plug-in micro set-top box.

[0086] Then, when the user initiates the device power-on action, the power-on action is treated as a second user action. In response to the detection of the second user action, the target streaming media is retrieved for the user to choose to play, and seamless playback is achieved simultaneously. The media is directly retrieved from the cache, reducing real-time processing latency.

[0087] Furthermore, the central computing platform monitors its own load status in real time and adjusts task allocation strategies to ensure response speed under concurrent requests from multiple devices.

[0088] Through the above technical solutions, this application effectively solves the problem of content adaptive processing efficiency in multi-screen interactive scenarios, realizes seamless connection of streaming media during switching, and uses cloud collaboration to process calculations, reducing the local computing load of plug-in micro set-top boxes, avoiding screen stuttering caused by insufficient hardware performance, reducing the amount of real-time transcoding tasks after prediction, shortening the switching response time to below the perceptible range, and improving the resource scheduling strategy to enhance the resource utilization of the central computing platform and ensure service stability in high-concurrency scenarios.

[0089] Example 2

[0090] Based on Example 1, please refer to the following: Figure 1 This further illustrates a specific embodiment of the adaptive processing method for multi-screen interactive content data in plug-in micro set-top boxes.

[0091] Step S1: Obtain a prediction model for the target plug-in mini set-top box based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in mini set-top box is configured such that if the user streams the same playback content from one plug-in mini set-top box to another plug-in mini set-top box, then the other plug-in mini set-top box is selected as the target plug-in mini set-top box.

[0092] Step S2: In response to detecting the first user action of the target user, obtain the current playback behavior.

[0093] Step S3: Based on the current playback behavior and the prediction model, obtain the target plug-in micro set-top box.

[0094] Step S31: Based on the current playback behavior, obtain the target streaming media data and send it to the target plug-in micro set-top box.

[0095] Step S32: The target plug-in micro set-top box is further configured to: cache target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for the target user to choose to play.

[0096] In this embodiment, there is a preparation step before step S1: the capability parameters of the plug-in micro set-top box and the devices connected to it are obtained through the plug-in micro set-top box and sent to the central computing platform; the plug-in micro set-top box identifier of each plug-in micro set-top box is obtained, and the capability parameters of the devices are stored in the capability parameter library and updated periodically.

[0097] When implementing the embodiments of this application, please refer to... Figure 2 In this embodiment, the central computing platform and the plug-in micro set-top boxes form a star topology. Each central computing platform is responsible for the computing of multiple plug-in micro set-top boxes within a region. The central computing platform and multiple plug-in micro set-top boxes constitute a star topology. At the same time, a user can have one or more home addresses, each home address is within the same central computing platform, and each home address has one or more set-top boxes. Each plug-in micro set-top box is configured with a different device according to different users' preferences. For example, the central computing platform corresponds to multiple plug-in micro set-top boxes, and each plug-in micro set-top box is configured with a different device, which can be any of the following: a television, a projector, a tablet computer, etc.

[0098] The plug-in mini set-top box collects its own hardware and real-time status information on itself and the various devices connected to it, including but not limited to: screen size, screen resolution, supported video encoding types, current network bandwidth, etc., and then uploads it to the central computing platform.

[0099] The central computing platform receives the capability parameters transmitted from the plug-in micro set-top boxes, generates a unique plug-in micro set-top box identifier for each box, and assembles the capability parameters collected by each box into a capability feature vector. It can be represented as , where i is the identifier for this plug-in mini set-top box. For resolution, For the list of supported codecs, For current network bandwidth, For example, the type of equipment.

[0100] Then, the identifier of each plug-in micro set-top box and its corresponding capability feature vector are returned to the central computing platform. This feature vector is stored in the capability parameter library. The plug-in micro set-top box periodically collects the capability parameters of the corresponding connected devices, thereby updating the capability parameter library of the central computing platform.

[0101] In step S1, a prediction model for the target plug-in mini set-top box is obtained based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in mini set-top box is configured such that if the user streams the same playback content from one plug-in mini set-top box to another plug-in mini set-top box, then the other plug-in mini set-top box is selected as the target plug-in mini set-top box.

[0102] In one possible implementation, step S1 includes:

[0103] Based on the target user's historical playback behavior, the probability of the user switching from the currently playing content to another target plug-in micro set-top box is obtained, in order to construct a predictive model for the target plug-in micro set-top box.

[0104] Based on the target user's historical playback behavior, playback device data and playback content data are time-series processed to obtain user usage logs;

[0105] Based on user logs, the probability of a user switching the currently playing content to another target plug-in mini set-top box is obtained; where the switching probability represents the likelihood that a user will switch the currently playing content from the current plug-in mini set-top box to the target plug-in mini set-top box;

[0106] Construct a training dataset based on the switching probabilities;

[0107] Based on user logs and the Transformer model, a time series prediction model is obtained.

[0108] Use user usage logs as input to output the switching probability of each potential target plug-in micro set-top box;

[0109] Based on historical playback behavior and training dataset, a time series prediction model is trained to minimize the cross-entropy loss between the prediction probability of the time series prediction model based on historical playback behavior and the real training dataset; wherein, when the output switching probability is higher than a preset threshold, the currently playing plug-in micro set-top box and the predicted plug-in micro set-top box after switching are obtained, and the plug-in micro set-top box after switching is used as the target plug-in micro set-top box for prediction.

[0110] The time series prediction model that minimizes cross-entropy loss is used as the prediction model for the target plug-in micro set-top box.

[0111] Based on the current streaming media being played on the device connected to the current plug-in mini set-top box, obtain the target streaming media for the device connected to the target plug-in mini set-top box.

[0112] The target streaming media will be adapted and stored in the cache area of ​​the corresponding target plug-in mini set-top box.

[0113] In one possible implementation, a predictive model of the target plug-in set-top box is obtained based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in set-top box is configured to use the other plug-in set-top box as the target plug-in set-top box if the user transfers the same playback content from one playback plug-in set-top box to another, including:

[0114] Acquire historical user behavior and historical status information to obtain a time series of user behavior and device status of devices connected to the set-top box;

[0115] Based on the time series, determine the likelihood of a user switching;

[0116] Based on the possibility of switching, for all available devices connected to the plug-in mini set-top box, calculate the probability that the plug-in mini set-top box becomes the target plug-in mini set-top box, obtain the probability distribution of the current playback content being transferred to the target plug-in mini set-top box, and obtain the prediction model of the target plug-in mini set-top box.

[0117] In implementing this application, we considered further optimizing the multi-screen interactive user experience, and thus introduced a predictive preprocessing and pre-distribution mechanism to reduce streaming switching latency. The central computing platform continuously monitors the device through a plug-in micro set-top box, then analyzes user behavior patterns and device usage habits, and combines predictive algorithms to predict future user interactions with multi-screen content.

[0118] The plug-in mini set-top box is responsible for collecting historical playback behavior and transmitting it to a central computing platform for analysis. This platform generates user logs across different devices, including: viewing history (e.g., which devices the user uses to watch specific types of content); interaction patterns (e.g., whether the user frequently switches between tablets or other TVs while watching TV); viewing duration (e.g., how long the user watched before switching); content type preferences (e.g., watching movies on TV, watching live sports on a tablet, and watching documentaries on other TVs); and viewing time (e.g., using a tablet during the day and TV at night). These user logs are collected and stored in real time.

[0119] Then, a time-series-based prediction model for the target plug-in micro set-top box is established using the Transformer model to predict the likelihood of a user switching at a future point in time and the most probable target plug-in micro set-top box. The Transformer model, due to its self-attention mechanism, can effectively capture long-range dependencies and complex patterns in the sequence, thus performing well in sequence prediction tasks. The input to the prediction model for the target plug-in micro set-top box is a time-series-based user usage log and various real-time capability feature vectors. For example, each element in the current state sequence This can include: the identifier of the currently used plug-in mini set-top box, the content identifier of the streaming media being played, time, day of the week, date, the real-time network bandwidth of the current plug-in mini set-top box, battery level, etc., for a total of N states. The model output is: for each potential target plug-in mini set-top box j, the corresponding capability feature vector is denoted as... The user will transfer the current content to the target plug-in mini set-top box. The probability of.

[0120] The prediction model for the target plug-in micro set-top box is trained using a large amount of historical user behavior data from user logs. The final model minimizes the cross-entropy loss between the predicted probability distribution and the actual user behavior to obtain the prediction. Therefore, the Softmax function is used to output the probability distribution, expressed as:

[0121]

[0122] in, The Transformer model is used for the target plug-in micro set-top box. The predicted score, M is the number of all available plug-in micro set-top boxes. The Transformer model is used for the target plug-in micro set-top box. The predicted score, Sum the predicted scores for all available plug-in micro set-top boxes, in Figure 2 In the model, under home address 1, the corresponding available devices are plug-in micro set-top box 1 and plug-in micro set-top box 2, because there is no corresponding device for plug-in micro set-top box 3. The model is trained on the central computing platform to provide low-latency real-time prediction. After obtaining the prediction results, pre-adaptation and pre-caching are performed.

[0123] When the prediction model of the target plug-in mini set-top box outputs a high probability of a certain switch, for example, exceeding a preset threshold θ (e.g., θ=0.7), the content data of the currently playing streaming media is immediately adaptively processed and transformed into the target streaming media adapted to the device connected to the target plug-in mini set-top box, thus providing advance prediction for the target plug-in mini set-top box. The central computing platform generates adapted target streaming media. For example, if it predicts that a user might switch from a TV to a tablet that supports H264 / 1080p / SDR, it pre-encodes and scales the currently playing content, converting it into an H264 / 1080p / SDR version. This pre-processed target streaming media is then cached in a storage system on the central computing platform and pre-distributed to the cache of the target plug-in micro-set-top box.

[0124] The pre-distribution strategy can be adjusted based on geographic location. For example, a user's geographic location includes two home addresses. In home address 1, there is a plug-in mini set-top box 1, plug-in mini set-top box 2, and plug-in mini set-top box 3. Plug-in mini set-top box 1 corresponds to device A, and plug-in mini set-top box 2 corresponds to device B. However, in another home address 2, the user has another plug-in mini set-top box 4, which corresponds to device C.

[0125] On Monday, a user is watching on device B from home address 1. At this moment, the prediction model of the target plug-in mini set-top box on my central computing platform determines that the user is watching on home address 1 and outputs that the user will watch on device A. It then issues the probability of switching the current playback content stream from plug-in mini set-top box 2 to plug-in mini set-top box 1. If the probability is higher than the threshold, it will pre-encode and scale the current playback content, converting the current playback content into a version adapted to device A, and pre-distributing the pre-processed target streaming media adapted to device A into the buffer of plug-in mini set-top box 1.

[0126] If today is Thursday, and the user is watching on device B at home address 1, the prediction model of the target plug-in mini set-top box on the central computing platform determines that the user is watching on home address 2. It outputs that the user will watch on device C and issues a switching probability to transfer the current playback content of plug-in mini set-top box 2 to plug-in mini set-top box 4. If the probability is higher than the threshold, the current playback content will be transcoded and scaled in advance. The original streaming media of the current playback content is found and pre-distributed to the buffer of plug-in mini set-top box 4. Then, plug-in mini set-top box 4 converts it into a version adapted to device C and stores the pre-processed target streaming media adapted to device C in the buffer.

[0127] When a switch actually occurs, the plug-in mini set-top box detects the second user action of its own device powering on. Then, it can directly retrieve the adapted target streaming media from the plug-in mini set-top box's buffer. The plug-in mini set-top box can immediately distribute the corresponding target streaming media to the corresponding device and start playback immediately, thereby further reducing the delay in content delivery. This allows the corresponding plug-in mini set-top box to respond immediately when the user turns on the target device, with most or even all of the adaptation work already completed. The user can play directly after making a selection, greatly improving the switching response speed and user experience.

[0128] Streaming currently playing content from the currently playing plug-in micro set-top box to another plug-in micro set-top box includes streaming the current content to the home address of a predicted target plug-in micro set-top box, on at least one device's set-top box.

[0129] In one possible implementation, a predictive model for the target plug-in mini set-top box is obtained based on the target user's historical playback behavior, including:

[0130] Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the identifier of the plug-in mini set-top box; the user interaction event includes a third user action, the third user action including at least a user initiating a content playback request on any device; in response to detecting a third user action, receive the third user action sent by the currently playing plug-in mini set-top box and the identifier of the currently playing plug-in mini set-top box.

[0131] Based on the content playback request, obtain the original streaming media from the content source;

[0132] The central computing platform sends the original streaming media, based on the identifier of the currently playing plug-in mini set-top box, to the currently playing plug-in mini set-top box; where:

[0133] The currently playing plug-in mini set-top box performs adaptive content data processing based on the acquired current playback capability parameters to obtain a currently playing streaming media adapted to the currently playing plug-in mini set-top box; wherein, the adaptive content data processing includes:

[0134] When the resolution of the original streaming media is higher than the current playback capability parameters, the current playback plug-in mini set-top box obtains the current playback streaming media with a video resolution adapted based on the original streaming media and an image scaling algorithm.

[0135] When the encoding format of the original streaming media is not directly supported by the current playback device, the current playback plug-in mini set-top box converts the original streaming media according to its encoding format to obtain the current playback streaming media with the encoding format supported by the current playback device.

[0136] Based on the network bandwidth of the current playback plug-in mini set-top box, and using rate-distortion theory, the highest bitrate of the currently playing streaming media is obtained.

[0137] In a possible implementation of this embodiment, based on the network bandwidth of the currently playing plug-in micro set-top box and using rate-distortion theory, the highest bitrate of the currently playing streaming media is obtained, including:

[0138] Maximize streaming quality and minimize distortion within network bandwidth constraints;

[0139] Based on the original streaming media, compression is performed, and the pixel difference between the original streaming media and the compressed streaming media is calculated to obtain the distortion.

[0140] Based on the distortion degree and network bandwidth, and using rate-distortion theory, we obtain the rate-distortion theory.

[0141] Based on rate-distortion theory, the minimum quantization parameter is obtained through iterative bisection to obtain the current streaming media with the highest bitrate.

[0142] When this application embodiment is implemented, when a user interacts with the computer through any device and generates a user interaction event, and the user inputs a request to play certain content, the plug-in micro set-top box connected to the device immediately sends a content playback request.

[0143] The central computing platform receives status information and user interaction events from each plug-in micro set-top box; the status information includes the identifier of the plug-in micro set-top box; the user interaction events include third user actions, which at least include a user initiating a content playback request on any device.

[0144] In response to the detection of third-user behavior, the central computing platform receives the third-user behavior and the identifier of the currently playing plug-in micro set-top box sent by the current playback plug-in micro set-top box.

[0145] Then, when the central computing platform receives the content playback request from this plug-in micro set-top box, it sends the plug-in micro set-top box identifier and the content playback request to the central computing platform. The content playback request includes at least the specific requested media content, such as the user currently requesting to watch a specific variety show. Then, based on the content playback request, the central computing platform obtains the original streaming media based on the content source. The original streaming media is obtained by querying the content source based on the requested media content in the user's behavior. The content source includes at least the original streaming media and a preset corresponding media content identifier. The content identifier is used to query the content source to find the corresponding original streaming media. The original streaming media includes the original video, audio, subtitles, resolution, etc.

[0146] For example, a user's third user action is to select the variety show "XXX" on a portable flat-screen TV in the kitchen. The plug-in micro set-top box connected to this flat-screen TV is denoted as IMSTB-C. Based on the content identifiers of various streaming media within the content source, the content identifier of the variety show "XXX" is found to be variety show A. At this time, my content playback request is to play variety show A. The content playback event trigger signal is: variety show A, playback progress 00:00, IMSTB-C. Then, based on the content source, that is, the film and television library connected to the central computing platform, a search is performed to find the original streaming media of the variety show "XXX".

[0147] Finally, the central computing platform retrieves the latest capability feature vector from the capability parameter database based on the plug-in micro set-top box identifier. Then, the original streaming media and capability feature vector are returned to the plug-in mini set-top box. At this point, the plug-in mini set-top box no longer needs to collect data from the device it is connected to. Based on the latest capability feature vector, the plug-in mini set-top box makes a judgment on the original streaming media and performs simple adaptive content data processing, and then processes it into streaming media that is currently playing on the device connected to the plug-in mini set-top box.

[0148] In this embodiment, the currently playing plug-in mini set-top box performs adaptive content data processing based on the acquired current playback capability parameters to obtain the currently playing streaming media adapted to the currently playing plug-in mini set-top box; wherein, the adaptive content data processing includes:

[0149] First, based on the identifier of the plug-in mini set-top box, the capability feature vector of the currently playing plug-in mini set-top box is found in the capability parameter library. Then, the hardware capability parameters and display capability parameters of the devices connected to the currently playing plug-in mini set-top box are obtained, and the specific target plug-in mini set-top box capability parameters are found, including at least the highest resolution, supported encoding formats, and network bandwidth.

[0150] First, based on the hardware, find the most objective limitations of the current playback device. By determining the device's maximum resolution and supported encoding formats, the video format is determined, ensuring device compatibility.

[0151] When the resolution of the original streaming media is higher than the current playback capability parameters, based on the original streaming media and an image scaling algorithm, a new streaming media with a resolution adapted to the current playback device is obtained. For example, if the resolution of the current playback device requested by the user is 3840x2160 pixels, while the encoding format of our original streaming media is a professional-grade or lossless compression format, the original video resolution at this moment is 7680x4320 pixels. Obviously, the resolution of the original streaming media is higher than the current playback capability parameters, so we perform a downsampling operation and demultiplex the encoding format of the original streaming media through demultiplexing, decoding, image processing, and re-encoding into an encoding format supported by the current playback device.

[0152] In the implementation of this application embodiment, the capability parameters connected to the currently playing insert-type micro set-top box are obtained by searching in the capability parameter library based on the identifier of the insert-type micro set-top box.

[0153] When the resolution of the original streaming media is higher than the current playback capability parameters, the current playback streaming media with a video resolution adapted is obtained based on the original streaming media and an image scaling algorithm.

[0154] If the original video resolution is higher than the target display device's resolution, the plug-in mini set-top box will perform a downsampling operation. This invention employs a Lanczos-based resampling image algorithm, which can effectively preserve edges and provides good anti-aliasing. The Lanczos kernel function... Represented as:

[0155]

[0156] in, It's the sinc function. In signal processing, the sinc function is a commonly used function, also known as the sine interpolation function, and can be represented as: ; This refers to the size of the kernel function, used to adjust the size of the filtering window. It is typically set to 2 or 3; in practical applications, this embodiment uses... .

[0157] For each output pixel Its brightness value The convolution formula is used to calculate the neighborhood pixels of each frame of the input raw streaming media, and it is represented as:

[0158]

[0159] in, The input image in pixels The brightness value at that location; Scaling ratios between different devices, such as scaling from 8K resolution to 4K resolution by a scaling ratio of 2; For Lanczos kernel functions.

[0160] When the encoding format of the original streaming media is not directly supported by the current playback device, it is converted according to the encoding format of the original streaming media to obtain the current playback streaming media with an encoding format supported by the current playback device;

[0161] If the original video encoding format is not directly supported by the target display device or is inefficient, the original video is then converted to the most efficient encoding format supported by the current playback device. For example, for 4K / HDR content, the H.265 encoding format is chosen. This conversion process involves multi-stage bitstream processing.

[0162] During adaptive processing of content data, the configuration parameters are adjusted based on the latest capability feature vector of the currently playing plug-in micro set-top box. First, the original streaming media is separated into video track, audio track and subtitle track by a demultiplexer. Then, the video track is decompressed into raw pixel data by a decoder. Then, these pixel data are processed by image processing, such as color space conversion and HDR image processing technology. Finally, they are re-encoded into the target format by an efficient encoder.

[0163] Based on the network bandwidth of the current playback plug-in mini set-top box, and using rate-distortion theory, the highest bitrate of the currently playing streaming media is obtained.

[0164] In practice, after obtaining the highest bitrate of the currently playing streaming media that is compatible with the connected device of the currently playing plug-in mini set-top box, the user selects the appropriate audio track and subtitle track based on their language preference, whether they prefer Chinese or English.

[0165] In this embodiment, based on the network bandwidth of the current playback plug-in mini set-top box and using rate-distortion theory, the highest bitrate of the currently playing streaming media is obtained, including:

[0166] The central computing platform queries the capability parameter database to determine the device's maximum resolution and encoding format. This ensures compatibility between the video and the current playback device. Once compatibility is confirmed, the optimal playback quality for the streaming media can be determined. To achieve the best video quality within limited network bandwidth, the video quality is maximized based on the current network bandwidth of the playback set-top box. The constraint is that the size of the generated video file, i.e., the bitrate, cannot exceed the network bandwidth capacity of the playback set-top box. Then, the bitrate and image quality can be balanced based on rate-distortion theory.

[0167] Based on the currently playing streaming media whose current playback resolution and encoding format have been confirmed, and according to the capability feature vector of the current playback plug-in mini set-top box... Network bandwidth Under the constraint of network bandwidth, maximize the quality of the currently playing streaming media, minimize distortion, and find the optimal video bitrate that can adapt to the network bandwidth of the currently playing plug-in mini set-top box, as the current streaming media with the highest bitrate.

[0168] The process involves compressing the original streaming media, calculating the difference between each pixel in the original streaming media and the compressed media content, and obtaining the distortion level. The distortion level is a specific numerical value that quantifies the loss of image quality during compression. The smaller the distortion level value, the less distortion there is and the better the image quality.

[0169] Bitrate is the amount of video data transmitted per second. A higher bitrate generally results in better picture quality, but also larger file sizes. Bitrate directly affects video transmission quality, while bandwidth determines transmission capacity; the two are mutually restrictive.

[0170] Based on distortion and network bandwidth, and grounded in rate-distortion theory, we obtain the rate-distortion model. This model is essentially an objective function, expressed as: ;

[0171] in, It refers to distortion; the essence of the model is to find the minimum distortion rate. It's the bitrate, which is limited by network bandwidth. It is a Lagrange multiplier used to balance bit rate and distortion. The goal of rate-distortion theory is to minimize the bit rate of the data by giving an allowable level of distortion in advance. Then, through this theory, more accurate decisions can be made in multiple fields such as data compression, image compression, and speech coding to minimize distortion while reducing the number of bits required for encoding as much as possible.

[0172] Based on rate-distortion theory, an iterative binary search method can be used to determine the optimal QP value. The QP value adjusts the quantization parameter, quantizing the compression intensity. It's a parameter controlling quality and bitrate in video coding. A smaller QP value results in lower compression intensity, preserving more detail and providing better image quality to the user, but at a higher bitrate. Then, it can be used... Lagrange multipliers are used to determine image quality and bitrate. In a complex scene, in order to maintain image quality, the central computing platform will choose a larger λ, which will allow a slightly higher bitrate and thus use a smaller QP.

[0173] Then you can try a QP value to see if the output bitrate is higher or lower than the network bandwidth. If the value is too high, increase QP and try again; if it's too low, decrease QP and try again. Repeat this process, using the simplest binary search method for iterative iteration, to quickly approximate the network bandwidth. And less than the optimal QP for bandwidth.

[0174] Then, during the adaptive processing of content data, the current playback plug-in mini set-top box combines a two-pass encoding strategy. The first pass analyzes the video content and complexity, and the second pass controls the bitrate based on the analysis results. Both encoding strategies are relatively simple and do not consume too many resources.

[0175] During the first pass of encoding, the entire video file is quickly and at low quality read to analyze its content complexity. For example, fast-moving scenes and rich texture details require a higher bitrate to maintain quality, while static images and solid-color backgrounds only require a very low bitrate.

[0176] This allows for a quick second pass of encoding. Based on the bitrate complexity distribution map obtained from the first pass analysis, the encoder assigns different bitrates to segments of varying complexity during the second pass. Complex segments are given higher bitrates and smaller QPs, while simple segments are given lower bitrates and larger QPs.

[0177] This allows for maximizing overall visual quality without changing the total bitrate, avoiding wasting bitrate in simple scenarios.

[0178] In step S2, in response to detecting the first user action of the target user, the current playback action is obtained.

[0179] In one possible implementation, in response to detecting a first user action of the target user, the current playback action is obtained, including:

[0180] Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the content identifier of the currently playing streaming media, the interruption playback progress timestamp, and the identifier of the currently playing plug-in mini set-top box; the user interaction event includes a first user action, which includes at least a power-off action of the device connected to the plug-in mini set-top box turning off.

[0181] In response to the detection of the first user action, if the streaming media content being played at the time of the device power-off action has not yet finished playing, the plug-in micro set-top box will return the status information as a trigger signal for the content interruption event to the central computing platform.

[0182] In step S3, based on the status information and the currently playing streaming media, the original streaming media and the target plug-in micro set-top box are obtained according to the prediction model of the target plug-in micro set-top box.

[0183] In one possible implementation, step S3 includes:

[0184] In response to the detection of the first user behavior by the currently playing plug-in micro set-top box, if the device is currently playing streaming media content at the time of power-off and the content has not yet finished playing, the currently playing streaming media content will be switched from the currently playing plug-in micro set-top box to another plug-in micro set-top box for playback.

[0185] Based on the trigger signal, obtain the content identifier and the interruption playback progress timestamp;

[0186] Based on the current status information and the content identifier of the currently playing streaming media, the predicted target plug-in micro set-top box is obtained based on the prediction model of the target plug-in micro set-top box.

[0187] Based on the content identifier and the content source, obtain the original streaming media;

[0188] Based on the predicted target plug-in micro set-top box identifier, the original streaming media is pre-distributed to the target plug-in micro set-top box's buffer.

[0189] In step S31, the target streaming media data is obtained and sent to the target plug-in micro set-top box based on the current playback behavior.

[0190] The currently playing plug-in mini set-top box obtains the current playback capability parameters of the connected device and performs adaptive content data processing to obtain the target streaming media adapted to the target plug-in mini set-top box.

[0191] In one possible implementation, pre-distribution to the buffer of the target plug-in micro set-top box includes:

[0192] At least send to the buffer of the target plug-in mini set-top box;

[0193] And or respectively sent to the buffers of all plug-in mini set-top boxes under the home address where the target plug-in mini set-top box is located.

[0194] In the implementation of this application embodiment, when a user is watching through any device, the device being watched is the current playback device. At this time, a user interaction event occurs, which is when the user turns off the current playback device. The corresponding current playback plug-in micro set-top box detects the first user action and turns off. The central computing platform receives status information and user interaction events from each plug-in micro set-top box. The status information includes the content identifier of the currently playing streaming media, the interruption playback progress timestamp, and the identifier of the current playback plug-in micro set-top box. The user interaction event includes the first user action, which includes at least the shutdown action of the device connected to the plug-in micro set-top box turning off. In response to detecting the first user action, the central computing platform first determines whether the currently playing streaming media has finished playing. If the content of the currently playing streaming media at the time of the shutdown action has not yet finished playing, the plug-in micro set-top box returns the status information as a trigger signal for a content interruption event to the central computing platform.

[0195] Based on a predictive model of the target plug-in micro set-top box, the central computing platform transfers the currently playing content from the currently playing plug-in micro set-top box to the target plug-in micro set-top box.

[0196] In practice, content can be transferred to another target plug-in set-top box, or the current content can be transferred to the home address of the predicted target plug-in set-top box, at least one plug-in set-top box.

[0197] For example, if a user is watching a movie on the living room TV at 8:00 PM and plans to turn it off, then wash up and return to their bedroom to continue watching on their computer, to seamlessly switch the viewing experience to their tablet in the bedroom, the user interaction event of turning off the living room TV at 8:00 PM is treated as the first user action. The plug-in mini set-top box in the living room immediately receives this and sends the trigger signal for the content interruption event and the current status information to the central computing platform. The central computing platform responds by preprocessing the unfinished content upon detecting the first user action.

[0198] Each plug-in mini set-top box maintains time synchronization with the central computing platform, so its timestamp is also synchronized with the central computing platform. Upon receiving a content interruption event trigger signal, the central computing platform first re-accesses the content source using the media content identifier to obtain the original streaming media. Simultaneously, it acquires the capabilities of the target plug-in mini set-top box, retrieving the latest capability parameters from the capability parameter database. Then it returns to the target plug-in mini set-top box for adaptive processing of the content data. This adaptive processing is the same as step S2, except that the current playback is handled by the plug-in mini set-top box. It has been switched to the new target plug-in micro set-top box. Following the method in step S2, based on the latest capability feature vector of the target plug-in micro set-top box's real-time status and hardware conditions... The system performs real-time adaptive processing of streaming media. For example, if the original content is 4K HDR, the TV is 4K / HDR / H265, and the target tablet is 2K / SDR / H264, the central computing platform will perform 4K to 2K downsampling, HDR to SDR tone mapping, and encoding format conversion, changing from H265 encoding format to H264 encoding format. Then, the target plug-in micro set-top box will store the processed target streaming media and interrupted playback progress timestamps in the buffer.

[0199] Then the central computing platform synchronizes the playback status, including but not limited to: the current volume setting (the user's current volume setting is 80%), the selected audio track, the subtitle language (the user's current selection is Chinese), the current playback speed (1.0x), the last playback pause point, and the user's specific playback preferences. This playback status information is packaged and transmitted to the target plug-in micro set-top box.

[0200] The target plug-in mini set-top box stores the target streaming media, along with interrupted playback progress timestamps and synchronized playback status information. Upon receiving user interaction events, including second user actions and power-on actions, the target plug-in mini set-top box begins retrieving the target streaming media for the user to select for playback. After the user makes a selection, it can control the connected target device to seamlessly play the media starting from the specified playback timestamp, thus ensuring a continuous user experience. It also preloads the streaming media within a set time window, and this interrupted playback progress timestamp is within the time window, making it convenient for the user to select a time point to start playback.

[0201] In this process, the present invention centrally manages the playback status of all devices connected to the plug-in micro set-top box through a central computing platform and performs the above operations to achieve seamless switching. Then, the plug-in micro set-top box performs content adaptive processing, realizing multi-screen switching. When users switch content between different screens, they experience timely and continuous content playback. Under ideal network conditions, the switching latency can be controlled within the set milliseconds. Even if the user turns off the currently playing device and immediately turns on the target device, the user feels that the content is on the new screen in an instant, with almost no perceptible stuttering or interruption.

[0202] Step S32: The target plug-in micro set-top box is further configured to: cache target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for the target user to selectively play.

[0203] In one possible implementation, step S32 further includes:

[0204] Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the content identifier of the currently playing streaming media, the interruption playback progress timestamp, and the identifier of the currently playing plug-in mini set-top box; the user interaction events include a second user action, which includes at least the power-on action of any device.

[0205] When the first user action occurs, the original media stream is obtained based on the content identifier and pre-distributed to the buffer of the target plug-in micro set-top box;

[0206] The target plug-in micro set-top box receives the raw streaming media distributed from the central computing platform, performs adaptive content data processing to obtain the target media stream, stores it in its local buffer, and establishes a preload queue.

[0207] The preload queue includes one or more target streaming media entries waiting in the queue. Each target streaming media entry includes the content identifier of the corresponding target streaming media, the start timestamp of preloading, and a corresponding target streaming media stored in the buffer area of ​​the target plug-in micro set-top box.

[0208] The number of target streaming media entries waiting in the preload queue does not exceed a preset maximum value. If the preset maximum value is reached, the preload queue is full.

[0209] When a new target streaming media is received and the preload queue is full, the target plug-in mini set-top box processes it according to a preset replacement strategy.

[0210] The target plug-in mini set-top box is also configured to: in response to detecting a second user action and retrieving the target streaming media, the target plug-in mini set-top box immediately transmits the preload queue to the target device connected to it for the user to select for playback.

[0211] In one possible implementation, when a new target streaming media is received and the preload queue is full, the target plug-in micro set-top box processes the data according to a preset replacement strategy, including:

[0212] The replacement strategy is to start an independent timer for each target streaming media in the preload queue;

[0213] If a target streaming media in a preloaded queue is not accessed by a user within a preset time limit after being received, or if no deletion instruction is received from the central computing platform, the target plug-in micro set-top box will automatically delete the queue and the corresponding data stored in the buffer.

[0214] In one possible implementation, the target plug-in micro set-top box is further configured to: cache target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for optional playback by the target user, and further include:

[0215] When a user selects the target streaming media for a previously interrupted content event on the predicted target plug-in mini set-top box, the target plug-in mini set-top box reads data from the buffer and immediately resumes playback from the interrupted playback progress timestamp.

[0216] When the target plug-in micro set-top box successfully resumes playback, it returns a playback resume success signal to the central computing platform. The playback resume success signal includes at least the content identifier of the target streaming media and the identifier of the target plug-in micro set-top box.

[0217] After receiving the successful playback resume signal, the central computing platform generates a preloaded content deletion instruction and sends it to all target plug-in micro set-top boxes except those that are resuming playback. The target plug-in micro set-top boxes except those that are resuming playback are configured to have target streaming media entries in their local preload queues, which include the content identifier of the target streaming media for which playback is resuming.

[0218] Other target plug-in micro set-top boxes that receive a preload content deletion instruction immediately delete the corresponding target streaming media entry and its data in the buffer from their preload queue.

[0219] In the implementation of this application embodiment, after each target plug-in micro set-top box receives the target streaming media pushed from the central computing platform, its internal local preloading queue processes it. First, it is stored in the local content data buffer of the target plug-in micro set-top box. Each received preloaded target streaming media creates a target streaming media entry in the preloading queue. The entry is a data structure that includes a content identifier, the start timestamp of the preloaded data (which is the starting position of the preloaded streaming media in the original content), and the reception time of the target streaming media. Then, a strict queue management replacement strategy is executed to balance the utilization efficiency of storage resources and user experience requirements.

[0220] First, ensure that the number of streaming media items waiting in the preload queue does not exceed the preset maximum value, such as 6 streaming media items. When a new target streaming media is received and the queue is full, trigger the replacement mechanism.

[0221] The replacement strategy doesn't simply remove the oldest or smallest data; instead, it makes decisions based on a priority evaluation function. This function comprehensively considers factors such as the frequency with which users have accessed this type of content on this device in the past, and even the remaining playback time of the content. For example, the existing queue entry with the lowest priority score and its associated data will be removed to make room for new content. If multiple entries have the same priority score, the earliest preloaded entry will be removed.

[0222] Furthermore, an independent timer is started for each target streaming media entry. To prevent storage resources from being occupied by unused preloaded data for a long time, if a preloaded content is not accessed by a user or does not receive a deletion instruction from the central platform within a preset time limit after being received, such as 10 days, the queue entry and its corresponding data stored in the local content data buffer will be automatically deleted, thereby releasing storage resources.

[0223] However, the timing period can also be remotely configured and dynamically adjusted by the central computing platform. For example, during periods of low user activity at night, the time limit can be appropriately shortened to release resources more quickly; while during periods of high user activity, such as weekends, the time limit can be extended to provide a longer preloading window and improve the user experience.

[0224] This embodiment also includes step S4, which, based on the computing, storage and network resources of the central computing platform, and using a resource scheduling strategy, obtains a scheduling strategy that minimizes the average processing latency of all user behaviors while maximizing resource utilization through an objective function.

[0225] The central computing platform schedules its internal computing, storage, and network resources. Based on resource scheduling strategies and through an objective function, it minimizes the average processing latency of all user behaviors while maximizing resource utilization. This ensures that the platform can meet the adaptive processing needs of content data in high-concurrency, real-time scenarios with multiple plug-in micro set-top boxes.

[0226] The central computing platform dynamically schedules and optimizes the computing, storage, and network resources within the platform. The computing within the platform includes at least CPU and GPU to ensure high-concurrency, real-time adaptive processing of content.

[0227] Because the central computing platform has various functions, including video transcoding, resolution scaling, bitrate control, and HDR / SDR conversion, with no overlap between them, containerization technology is used to manage these functions. First, different functions are encapsulated as microservices, each running in an independent container. When there is a surge in real-time content requests or predictive preprocessing tasks, such as during prime time from 7:00 PM to 10:00 PM when most users are watching media on different devices, the central computing platform can automatically adjust its load balancing strategy and resource utilization thresholds. This includes setting round-robin and least-connection load balancing strategies, and triggering alerts when CPU utilization exceeds 70%, automatically scaling up the number of containers for related services. For example, if the CPU utilization of the real-time transcoding service is too high, the central computing platform will automatically start more transcoding containers and distribute new transcoding requests to these new containers.

[0228] The central computing platform continuously monitors its CPU utilization, memory usage, and GPU load. Based on resource scheduling algorithms, high-priority tasks, such as user-initiated real-time streaming switching tasks, will receive higher computing resource allocation and shorter scheduling cycles to ensure that their response time is minimized. In contrast, predictive preprocessing tasks can be executed when the load is low as background tasks to avoid affecting real-time services.

[0229] The objective function of the resource scheduling algorithm is set to minimize the average processing latency of all user actions while maximizing resource utilization, which can be expressed as:

[0230]

[0231] in, It is the processing delay of the i-th request, for example, transcoding time or switching time. It is its weight, for example, the weight of real-time task switching. Much higher than preprocessing tasks, , , The resource utilization rates are the current utilization rates of CPU, GPU, and memory, respectively. It is a preset maximum utilization threshold, so that the central computing platform can achieve optimal central computing platform performance by continuously adjusting task priorities and resource allocation.

[0232] Furthermore, in addition to reporting standard hardware capabilities, the plug-in micro set-top box periodically uploads real-time ambient light information for its connected devices and the user's viewing distance from their current position. Ambient light information can be obtained through an illuminance sensor integrated near the plug-in micro set-top box or monitor. Viewing distance information can be obtained by the user manually inputting their typical viewing distance in the application; alternatively, it can be measured in real-time using a depth camera integrated into the plug-in micro set-top box; or, more advancedly, estimated using an integrated camera combined with facial recognition and human pose estimation technology, provided the device supports it, such as a television. The central computing platform incorporates these environmental parameters into its content adaptive processing. For example, when strong ambient light is detected, the contrast or brightness of the video stream can be adjusted to improve viewing clarity and detail visibility. At longer viewing distances, such as over 3 meters, the font size of subtitles or the scaling of interface elements in the video can be appropriately increased to optimize the user experience and ensure readability. These adjustment parameters, such as brightness factors, contrast enhancement, and font scaling, can be predefined and stored in configuration files or predicted in real-time by machine learning models based on environmental input.

[0233] The data processing method for multi-screen interaction based on a plug-in micro set-top box provided by this invention completely changes the role of the traditional plug-in micro set-top box in multi-screen interaction through the above series of steps. The plug-in micro set-top box no longer undertakes the heavy and resource-intensive content adaptation task, but only serves as a lightweight data collection, instruction uploading and pre-adaptation stream reception, and distributes it to the display terminal. All complex and high-load computing tasks are performed on a resource-rich central computing platform, thereby eliminating the multi-screen interaction performance bottleneck caused by the hardware performance limitations of the plug-in micro set-top box itself in the traditional solution.

[0234] Example 3

[0235] This is the third embodiment of the present invention. Based on embodiments 1 and 2, please refer to the following references. Figure 3 This is a schematic diagram of the structure of the plug-in micro set-top box multi-screen interactive content data adaptive processing system provided in this embodiment of the invention. This embodiment provides a plug-in micro set-top box multi-screen interactive content data adaptive processing system, including a prediction unit, a multi-screen interaction unit, and a content data processing unit that are electrically connected.

[0236] The prediction unit is configured to: obtain a prediction model of the target plug-in mini set-top box based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in mini set-top box is configured to select the other plug-in mini set-top box as the target plug-in mini set-top box if the user streams the same playback content from one playback plug-in mini set-top box to another plug-in mini set-top box.

[0237] The multi-screen interaction unit is configured to: obtain the current playback behavior in response to the detection of the first user behavior of the target user.

[0238] The content data processing unit is configured to: obtain the target streaming media adapted to the target plug-in micro set-top box based on the status information and the currently playing streaming media, using the prediction model of the target plug-in micro set-top box.

[0239] The multi-screen interaction unit is also configured to: in response to detecting a second user action, retrieve the target streaming media for the user to select for playback.

[0240] In the implementation of this application embodiment, it also includes a status information acquisition unit and a resource utilization optimization unit electrically connected to the multi-screen interaction unit.

[0241] The status information acquisition unit obtains the hardware parameters and real-time status data of the connected device through the communication interface, and transmits the acquired data to the central computing platform through the plug-in micro set-top box to establish a dynamically updated capability parameter library, providing basic parameters for streaming media adaptation.

[0242] The resource utilization optimization unit performs calculations on the resources of the central computing platform in the cloud. It adopts load balancing and resource scheduling strategies to optimize the allocation of central computing platform resources according to the needs of concurrent tasks, thereby ensuring processing efficiency.

[0243] The central computing platform's status information acquisition unit receives real-time status information and user interaction event data transmitted from the plug-in micro set-top box, using this as the basis for the prediction unit's predictions. The prediction unit builds a model based on historical data and identifies the target plug-in micro set-top box based on real-time data. The content data processing unit, based on the data from the status information acquisition unit and the target plug-in micro set-top box from the prediction unit, adaptively processes the streaming media and transmits the processed media to the multi-screen interaction unit. The multi-screen interaction unit responds to user interaction events, detecting different user behaviors, including power-on, power-off, and playback requests. It reacts according to specific user behaviors and feeds back to the central computing platform's content data processing unit for managing the pre-loaded queue. The resource utilization optimization unit receives data from the status information acquisition unit regarding the central computing platform's own computing resources, performs calculations, and adjusts the priority and resource allocation of the content data processing unit's calculation process to achieve optimal central computing platform performance.

[0244] Example 4

[0245] The fourth embodiment of the present invention differs from the previous embodiments in that:

[0246] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0247] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, or may be electrical, mechanical, or other forms of connection.

[0248] The units described as separate components may or may not be physically separate. As will be apparent to those skilled in the art, the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0249] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0250] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device, such as a personal computer, server, or grid device, to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0251] The above specific embodiments further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for adaptive processing of multi-screen interactive content data in a plug-in mini set-top box, characterized in that: include: Based on the target user's historical playback behavior, a predictive model for the target plug-in mini set-top box is obtained; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in mini set-top box is configured such that if the user streams the same playback content from one plug-in mini set-top box to another plug-in mini set-top box, then the other plug-in mini set-top box is selected as the target plug-in mini set-top box. In response to the detection of the first user action of the target user, the current playback behavior is obtained; The first user action includes at least the power-off action of turning off the device connected to the plug-in micro set-top box; Based on the current playback behavior and the prediction model, the target plug-in mini set-top box is obtained; Based on the current playback behavior, the target streaming media data is obtained and sent to the target plug-in micro set-top box; The target plug-in micro set-top box is further configured to: cache the target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for the target user to selectively play; the second user action includes at least the power-on action of any device; It is also equipped with a central computing platform, which is configured to: in response to detecting third user behavior, receive the third user behavior and the identifier of the currently playing plug-in micro set-top box sent by the current playback micro set-top box; When the central computing platform receives a content playback request from any plug-in micro set-top box, the identifier of the plug-in micro set-top box and the content playback request are sent to the central computing platform; the third user action includes at least the user initiating a content playback request on any device; the central computing platform finds the latest capability feature vector from the capability parameter library based on the identifier of the plug-in micro set-top box, and then returns the original streaming media and the capability feature vector together to the plug-in micro set-top box; then the plug-in micro set-top box obtains the target streaming media data adapted to the device connected to the plug-in micro set-top box based on the latest capability feature vector; The plug-in micro set-top box collects the hardware and real-time status information of itself and the various devices connected to it, and then uploads it to the central computing platform. The central computing platform receives the capability parameters transmitted from the plug-in micro set-top box, generates a unique plug-in micro set-top box identifier for each plug-in micro set-top box, and assembles the capability parameters collected by this plug-in micro set-top box into a capability feature vector. The capability feature vector includes at least resolution, current network bandwidth, and device type.

2. The adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box according to claim 1, characterized in that, The step of obtaining a prediction model for the target plug-in mini set-top box based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in mini set-top box is configured such that if a user switches the same playback content from one plug-in mini set-top box to another, the other plug-in mini set-top box stream is used as the target plug-in mini set-top box, including: Based on the playback device data and playback content data of the target user's historical playback behavior, time-series data is generated to obtain the user usage log. Based on the user usage logs, the switching probability of a user switching the currently playing content to another target plug-in mini set-top box is obtained; wherein, the switching probability represents the likelihood that a user will switch the currently playing content from the current plug-in mini set-top box to the target plug-in mini set-top box; Based on the switching probabilities, construct a training dataset; Based on the user usage logs, a time series prediction model is obtained using the Transformer model. Use user usage logs as input to output the switching probability of each potential target plug-in micro set-top box; The time series prediction model is trained based on the historical playback behavior and the training dataset to minimize the cross-entropy loss between the prediction probability of the time series prediction model based on the historical playback behavior and the real training dataset. The time series prediction model that minimizes cross-entropy loss is used as the prediction model for the target plug-in micro set-top box.

3. The adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box according to claim 2, characterized in that, The method of obtaining a prediction model for the target plug-in mini set-top box based on the target user's historical playback behavior further includes: Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the identifier of the currently playing plug-in mini set-top box; the user interaction event includes a third user action; in response to detecting a third user action, receive the third user action and the identifier of the currently playing plug-in mini set-top box sent by the currently playing plug-in mini set-top box. Based on the content playback request, the original streaming media is obtained from the content source; The original streaming media is sent to the currently playing plug-in mini set-top box based on the identifier of the currently playing plug-in mini set-top box; wherein: The currently playing plug-in mini set-top box performs adaptive content data processing based on the acquired current playback capability parameters to obtain a currently playing streaming media adapted to the currently playing plug-in mini set-top box; wherein, the adaptive content data processing includes: When the resolution of the original streaming media is higher than the current playback capability parameter, the current playback plug-in mini set-top box obtains the current playback streaming media with video resolution adapted based on the original streaming media and an image scaling algorithm. When the encoding format of the original streaming media is not directly supported by the current playback device, the current playback plug-in mini set-top box converts the original streaming media according to its encoding format to obtain the current playback streaming media with the encoding format supported by the current playback device. Based on the network bandwidth of the currently playing plug-in mini set-top box, and using rate-distortion theory, the current streaming media with the highest bitrate is obtained. The step of obtaining the highest bitrate current streaming media based on the network bandwidth of the currently playing plug-in micro set-top box, using rate-distortion theory, includes: Maximize streaming quality and minimize distortion within network bandwidth constraints; Based on the original streaming media, compression is performed, and the pixel difference between the original streaming media and the compressed streaming media is calculated to obtain the distortion. Based on the aforementioned distortion and network bandwidth, and using rate-distortion theory, the rate-distortion theory is obtained. Based on the rate-distortion theory, the minimum quantization parameter is obtained by iteratively performing a binary search method to obtain the current playback stream with the highest bitrate.

4. The adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box according to claim 1, characterized in that, The step of obtaining the current playback behavior in response to detecting a first user action of the target user includes: Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the content identifier of the currently playing streaming media, the interruption playback progress timestamp, and the identifier of the currently playing plug-in mini set-top box; the user interaction events include a first user action; In response to the detection of the first user action, if the streaming media content being played at the time of the power-off action has not yet been completed, the plug-in micro set-top box will use the status information as a trigger signal for a content interruption event and return it to the central computing platform.

5. The adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box according to claim 4, characterized in that, The step of obtaining the target plug-in mini set-top box based on the prediction model according to the current playback behavior; wherein: obtaining target streaming media data and sending it to the target plug-in mini set-top box according to the current playback behavior includes: In response to the current playback of the plug-in mini set-top box detecting a first user action; If the device is currently playing streaming media content at the time of power-off, and the content has not yet finished playing, then the currently playing streaming media content will be switched from the current playback plug-in micro set-top box to another plug-in micro set-top box for playback; Based on the trigger signal, obtain the content identifier and the interruption playback progress timestamp; Based on the current status information and the content identifier of the currently playing streaming media, the predicted target plug-in micro set-top box is obtained based on the prediction model of the target plug-in micro set-top box. Based on the content identifier, the original streaming media is obtained from the content source; Based on the predicted target plug-in micro set-top box identifier, the original streaming media is pre-distributed to the buffer of the target plug-in micro set-top box; The currently playing plug-in mini set-top box obtains the currently connected playback capability parameters and performs adaptive content data processing to obtain a target streaming media adapted to the target plug-in mini set-top box.

6. The adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box according to claim 5, characterized in that, The buffer area for pre-distributing to the target plug-in micro set-top box includes: At least sent to the buffer of the target plug-in micro set-top box; And / or respectively sent to the cache areas of all plug-in micro set-top boxes under the home address where the target plug-in micro set-top box is located.

7. The adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box according to claim 1, characterized in that, The target plug-in micro set-top box is further configured to: cache the target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for the target user to selectively play, including: Receive status information and user interaction events from each plug-in mini set-top box; wherein, the status information includes the content identifier of the currently playing streaming media, the interruption playback progress timestamp, and the identifier of the currently playing plug-in mini set-top box; the user interaction events include a second user action; When the original media stream is obtained based on the content identifier, it is pre-distributed to the buffer of the target plug-in micro set-top box; The target plug-in micro set-top box receives the original media stream distributed from the central computing platform, performs adaptive content data processing, obtains the target streaming media, stores it in its buffer, and establishes a preloading queue. The preload queue includes one or more target streaming media entries waiting in the queue. Each target streaming media entry includes the content identifier of the corresponding target streaming media, the start timestamp of preloading, and a corresponding target streaming media stored in the buffer area of ​​the target plug-in micro set-top box. The number of target streaming media entries waiting in the preloading queue does not exceed a preset maximum value. If the preset maximum value is reached, the preloading queue is full. When a new target streaming media is received and the preload queue is full, the target plug-in micro set-top box processes it according to a preset replacement strategy. The target plug-in micro set-top box is further configured to: in response to detecting a second user action, retrieve the target streaming media, and immediately transmit the preload queue to the target device connected to it for the user to select and play.

8. The adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box according to claim 7, characterized in that, When a new target streaming media is received, and the preloading queue is full, the target plug-in micro set-top box processes the data according to a preset replacement strategy, including: The replacement strategy involves starting an independent timer for each target streaming media in the preload queue; If a target streaming media in a preloaded queue is not accessed by a user within a preset time limit after being received, or if no deletion instruction is received from the central computing platform, the target plug-in micro set-top box will automatically delete the queue and its corresponding data stored in the buffer.

9. The adaptive processing method for multi-screen interactive content data of the plug-in micro set-top box according to claim 7, characterized in that, The target plug-in micro set-top box is further configured to: cache the target streaming media data, and in response to detecting a second user action, retrieve the target streaming media data for the target user to selectively play, further comprising: When a user selects the target streaming media for a previously interrupted content event on the predicted target plug-in mini set-top box, the target plug-in mini set-top box reads data from the buffer and immediately resumes playback from the interrupted playback progress timestamp. When the target plug-in micro set-top box successfully resumes playback, it returns a playback resume success signal to the central computing platform. The playback resume success signal includes at least the content identifier of the target streaming media to be resumed and the identifier of the target plug-in micro set-top box to be resumed. After receiving the playback resume success signal, the central computing platform generates a preload content deletion instruction and sends it to all target plug-in micro set-top boxes except those that are resuming playback; wherein, the target plug-in micro set-top boxes except those that are resuming playback are configured as follows: the target streaming media entry in their local preload queue includes the content identifier of the target streaming media for which playback is resuming. Other target plug-in micro set-top boxes that receive the preload content deletion instruction immediately delete the corresponding target streaming media entry and its data in the buffer from their preload queue.

10. A plug-in micro set-top box multi-screen interactive content data adaptive processing system, characterized in that, It includes a prediction unit with electrical connectivity, a multi-screen interaction unit, a content data processing unit, and a central computing platform; The prediction unit is configured to: obtain a prediction model of the target plug-in micro set-top box based on the target user's historical playback behavior; the playback behavior is configured to include at least playback device data and playback content data; the target plug-in micro set-top box is configured to use the other plug-in micro set-top box as the target plug-in micro set-top box if the user transfers the same playback content from one playback plug-in micro set-top box to another plug-in micro set-top box. The multi-screen interaction unit is configured to: obtain the current playback behavior in response to detecting the first user behavior of the target user; The first user action includes at least the power-off action of turning off the device connected to the plug-in micro set-top box; The content data processing unit is configured to: obtain a target plug-in mini set-top box based on the prediction model according to the current playback behavior; wherein: the target streaming media data is obtained according to the current playback behavior and sent to the target plug-in mini set-top box; The multi-screen interaction unit is further configured to: cache the target plug-in micro set-top box, and in response to detecting a second user action, retrieve the target streaming media data for the target user to selectively play; the second user action includes at least the power-on action of any device; It is also equipped with a central computing platform, which is configured to: in response to detecting a third user action, receive the third user action and the identifier of the currently playing plug-in micro set-top box sent by the current playback plug-in micro set-top box; when the central computing platform receives a content playback request from any plug-in micro set-top box, the identifier of any plug-in micro set-top box and the content playback request are sent to the central computing platform; the third user action includes at least a user initiating a content playback request on any device; the central computing platform finds the latest capability feature vector from the capability parameter library based on the identifier of the any plug-in micro set-top box, and then returns the original streaming media and the capability feature vector together to the any plug-in micro set-top box; then the any plug-in micro set-top box obtains the target streaming media data adapted to the device connected to the any plug-in micro set-top box based on the latest capability feature vector; The plug-in micro set-top box collects the hardware and real-time status information of itself and the various devices connected to it, and then uploads it to the central computing platform. The central computing platform receives the capability parameters transmitted from the plug-in micro set-top box, generates a unique plug-in micro set-top box identifier for each plug-in micro set-top box, and assembles the capability parameters collected by this plug-in micro set-top box into a capability feature vector. The capability feature vector includes at least resolution, current network bandwidth, and device type.

Images